How AI is changing health insurance underwriting and claims in NZ

AI in health insurance underwriting

AI already has a role in New Zealand health insurance underwriting, but only one of the five health insurers Policywise works with automates a decision, and only on straightforward applications.

Three of the five health insurers Policywise works with disclose using AI or automation somewhere in the application process.

One said in 2019 that it expected most applications through its online tool to be accepted in real time. Only one of the five describes an automated decision on a health application. Where the other four automate anything, it is scoped to non-health products.

In practice, AI mostly sits around the underwriting decision rather than in place of it. Two providers describe it:

• one provider's online application tool asks dynamic questions that adapt as you answer
• one provider lists summarising customer interactions and medical notes among its AI uses, and says its underwriters check an applicant's medical records to make the call

Where an insurer's tool does automate a decision, it is the straightforward, real-time acceptances. The tool's published claim is real-time acceptance, and nothing published describes an automated decline. The test is still what a prudent underwriter needs to know to decide whether to offer cover, and on what terms.

The upside for you is time and effort. The questions can adapt to your earlier answers, and a straightforward application can be accepted on the spot rather than days later. However, the risks sit in three places.

Bias

Health New Zealand warns that generative AI can reinforce social and health inequities because the data it learns from may carry those biases. It names Māori, women, older people, disabled people, and gender-diverse people among the groups poorly represented in that data.

That guidance covers the public health system rather than insurers, but New Zealand research on local healthcare data makes the same point.

Opacity

The Privacy Commissioner says even experts are often unable to explain how an AI tool reached a particular output. The Commissioner also cautions that a human in the loop may not be enough, because people overseeing automated systems can stop questioning them.

No AI-specific rule to fall back on

New Zealand has no AI Act. You are protected instead by the Privacy Act, which the Commissioner says applies to anyone using AI tools in New Zealand. Insurers have also owed a fair conduct duty since March 2025.

The Human Rights Act permits different terms by age, sex or disability only where the difference rests on actuarial or statistical data it is reasonable to rely on. For disability alone, if no such data exists, reputable medical or actuarial advice can be relied on instead, provided the treatment is reasonable in the circumstances.

What you disclose still matters, whatever the form looks like, and the duty is strict. As the law stands, you must tell your insurer anything that could affect its decision to insure you, and if you get that wrong, the insurer can cancel the policy back to the start.

From November 15, 2027 at the latest, the Contracts of Insurance Act 2024 replaces that duty with a requirement to take reasonable care not to make a misrepresentation. It also makes the remedies proportionate, so an honest mistake is treated differently from a deliberate one.

What an insurer may collect, and what it may do with it, should also be understood. Health insurers are health agencies under the Health Information Privacy Code, so they must collect only the health information they need.

The authority you sign can let it go past the form and ask your doctors, specialists, laboratories, and ACC for relevant records. Even so, a request for your complete notes should be the exception.

Insurers differ on whether that information can then train an AI model. Four of the five state a position, and the answers do not match.

One says your data is not used to train AI models. One says it may be. One says AI tools may use the data you enter to improve their predictions, with your personal information protected. One rules out external third-party models only.

Where an adviser comes in. An insurance adviser deals with the insurer on your behalf. They can tell you how a particular insurer underwrites a condition like yours before you apply, including whether you might face a stand-down period or a loading. They are also bound by statutory duties of ethical behaviour, competence and client care.

Examples of AI in health insurance underwriting

No New Zealand dispute about an AI-made underwriting decision has been published. The case studies the Insurance & Financial Services Ombudsman has published, across every category including health insurance, contain no complaint about an algorithmic or automated decision as at September 2026. The disputes that do reach it turn on people.

The cases that do reach it show what these disputes actually turn on:

• An adviser asked an insurer whether it wanted more from the client's doctor, and the insurer said it was not warranted. Cardiology investigations from 2015 surfaced at claim time years later, the trauma cover was cancelled from inception, and a 75% loading went on the life cover. IFSO did not uphold the complaint.
• In another complaint involving life insurance with pre-existing conditions, two independent underwriters were needed to settle whether undisclosed diabetes and high cholesterol would have changed the cover offered. That complaint was not upheld either.

Neither turned on technology. Both turned on what was asked, what was recorded, and who pushed back.

AI and pre-existing conditions

None of the five health insurers Policywise works with describes assessing your pre-existing conditions automatically, though one describes its underwriters checking an applicant's medical records.

A pre-existing condition exclusion can bite on a symptom, a sign, or a circumstance present when you took out the policy, rather than on a diagnosis. In one Ombudsman case, a knee claim was declined because records noted degenerative changes 15 years earlier. Nothing in that case turned on how the assessment was made.

Having pre-existing conditions does not mean you cannot get cover at all, or that these health issues will be automatically excluded from your policy. You can ask an insurer to cover a pre-existing condition, and if it agrees, you will probably pay a higher premium to reflect the risk. Or you may need to wait for a period before that part of your cover starts.

Switching insurers does not remove a pre-existing condition. A new insurer will typically re-assess it and is likely to treat it as pre-existing under its own policy. Policywise's guide to health insurance for pre-existing conditions goes through the options.

AI and health insurance claims

AI has barely surfaced in health insurance claims in New Zealand. Of the five health insurers Policywise works with, one discloses using AI-powered software to help assess the information in a claim. The other four say nothing about it either way, which is not the same as saying a person decides.

What the published material describes instead is electronic claiming rather than automated assessment. One insurer's affiliated provider network handles prior approval and claims online, and a separate insurer confirms claim and pre-approval outcomes with you by phone or email.

The upside. Electronic claiming takes paperwork out of the round trip between your specialist, the insurer, and you. Analysis of the New Zealand market describes AI triaging claims, automating routine ones and passing complex decisions to people, and sifting claims data for irregularities. That is commentary on what the technology can do, not evidence of what health insurers here are doing.

What still needs a person. The Privacy Commissioner expects human review built into decisions that carry real consequences. European supervisors draw the same line, treating a system that decides what a policyholder is paid as higher risk than one that retrieves documents. That is a European standard, not a New Zealand rule.

The risks. There is less to fall back on than you might assume. New Zealand has no equivalent of the European right not to be subject to a decision made solely by automated means. The Privacy Commissioner has said people should know why an automated decision was taken against them.

Public comfort with automated decisions is low. Only 35% of New Zealanders are comfortable with AI making automated financial decisions.

Screening at claim time cuts both ways. Assessing a claim can mean revisiting what you told the insurer when you applied, and material you did not disclose can cost you the policy rather than only the claim. Health information collected for a purpose can be used for that purpose without asking you again.

Insurers typically ask you to authorise this at application, which is what lets them approach your doctor later. The industry's Insurance Claims Register is a general-insurance facility, run by general insurers, not the health insurers on Policywise's panel.

Where an adviser comes in. An adviser arranges your policies with the insurer and may also be able to help when you claim. They can ask the insurer to explain how the policy applies to your situation

Talk to a Policywise adviser if a claim is not going the way you expected.

Chatbots, triage and faster service

Chatbots are a smaller part of the New Zealand health insurance claims journey than international coverage suggests. Two insurers launched AI assistants, one in 2017 and one in 2019, and there is no sign of either in service today. A third says it may use chatbots, and says that, wherever practicable, it will tell you when you are interacting with an AI-powered function.

One insurer has added an AI symptom checker to its member app, but that is clinical triage, not a claims tool. Lodging a claim or chasing its status today most likely means a person, an online form, or both.

A more immediate AI risk is the chatbot you use, not one the insurer uses. The Ombudsman warns that AI tools can oversimplify policy wording, miss exclusions, apply overseas rules to New Zealand policies, and invent past cases.

One AI summary it saw told a consumer that claim decisions are frequently overturned and that up to 80% to 90% of cases succeed if people persist. IFSO's own figure is that about one in four disputes it investigates succeed for the consumer.

What to do if an AI-assisted decision goes against you

If a health insurance application or claim goes against you, ask for the reason first. None of the five health insurers Policywise works with publicly promises to give you one.

The Ombudsman's advice is to ask the insurer to explain how your policy applies to your situation. The Privacy Act separately gives you rights over the information sitting behind the decision.

1. Start by asking for the reason, in writing: Ask which part of the policy, or which piece of information, the decision rests on.
2. You can also ask for your file: You can request the personal information an agency holds about you, and it must respond within 20 working days or tell you why it needs longer. An insurer can withhold confidential third-party opinion, but not simply because the file is its own staff's assessment of you. Whether the same applies to an AI-generated risk score has not been tested in New Zealand.
3. If something is wrong, ask for a correction: If something in the file is inaccurate, out of date, incomplete, or misleading, you can ask for it to be corrected. The agency has 20 working days to decide, and if it refuses, you can ask for a statement of correction to be attached and complain to the Privacy Commissioner. You cannot make an agency delete information it still needs.
4. If that doesn't resolve it, escalate: Use the insurer's own complaints process first. It has 2 months to resolve the complaint, and about 90% of cases that reach the Ombudsman are resolved by the provider along the way.

At deadlock, the Insurance & Financial Services Ombudsman scheme is free, independent, and can investigate claims up to $500,000 plus GST. All five health insurers Policywise works with belong to IFSO.

Some other NZ financial providers belong to Financial Services Complaints Limited instead. About one in four disputes the Ombudsman investigates ends well for the consumer.

Worried about how AI could affect your cover? Let Policywise help

Policywise is a 100% free service which tells you which health, life, and disability insurance provider best fits your needs. We offer fast, comprehensive, and easy-to-understand comparisons of all leading providers, and a simple summary clearly recommending which insurer is best for your situation.

Not all insurance policies are the same. Policywise can help you sort out the duds, avoid the lemons, understand the fine print and exclusions, and get the right insurance for you and your family.

We make the important decision of where to buy your insurance super easy. We’ll answer your questions, provide experienced advice and quotes, and manage all the back and forth throughout the application process. Taking out your cover through us means you'll have our lifetime support and claims advocacy, and we'll help you negotiate a positive outcome at claim time. We can also take care of lodging any claims on your behalf and back you up if the going gets tough.

Check out the reviews on our homepage for how other New Zealanders have found our service, because now is the time to get your personal, family, or business insurance sorted. Give your family or someone you love the most outstanding financial support possible. Book a 5-minute callback with Policywise today; our service is fast and free.